Re: [stella] 11 Invaders "Reloaded"

Subject: Re: [stella] 11 Invaders "Reloaded"
From: Thomas Jentzsch <tjentzsch@xxxxxx>
Date: Fri, 28 Nov 2003 00:04:47 +0100
On Thursday, November 27, 2003 at 23:06, Christopher Tumber wrote:
> You've overcome this problem using the illegal HMOVE trickery,
> however, so did Erik Mooney's old demo years ago. (In fact since
> Erik's works properly on emus and presumably on Kool-Aid man
> affected 2600's I'd suggest that you're not yet as far along as he
> got). When I did Space Instigators, it was only after I'd made the
> decision to drop to 9x5 that I realised Erik had found a way around
> it. And I think I asked Eckard about it or something and he
> explained the theory, but at the time since it was all undocumented
> and I'd only been working with the 2600 a few months I decided to
> just let it go.

Which demos are you talking about? I would like to have a look at

> The big trick is NOT displaying sprites. Specifically, leaving the
> space blank when aliens are killed off.

Yup, exactly.

> A big, big part of the Space Instigators code is to handle
> exceptions where you can no longer use the RESPn trick.

The only exceptions are when there is only one sprite 0 and/or 1 in
a row. There I am using the normal positioning code.

> However, at 11x5 the number of exceptions is increased
> exponentially and some of them may not be possible to do at all. For
> example:

> 10_0_0_0_01

> Is that even possible? The 0's would be RESPn trick, but is there
> a configuration of duplicated sprites that can span that vertical
> distance?

Did you have the chance to look at the demo I posted? It already
randomly switches the sprites on and off, and except for one case (one
sprite each, which I have corrected already) it works perfect.

And for the situation you are describing I can still use the RESPx
trick. NUSIx is usually set to two copies close (only except for
single sprites). So there is absolutely no problem, since as long as
you trigger it at least twice/row only the second copy is displayed.

> (BTW - All the previous stuff about GRPn is irrelevant, the
> reason there are so many different kernals is one for each vertical
> position of the formation. At the time I couldn't come up with any
> other delay loop with 1 cycle of accuracy (ie: a delay between 0 to
> 11 cycles or whatever)

I am using my beloved illegal opcodes for that (various length NOPs).

Have fun!
Thomas Jentzsch         | *** Every bit is sacred ! ***
tjentzsch at web dot de |

Archives (includes files) at
Unsub & more at

Current Thread